Two-dimensional digital image correlation-based microscopic failure type identification method for rock material numerical model

Abstract

The invention provides a method for solving transgranular or intergranular failure of a rock material numerical model based on two-dimensional digital image correlation, and belongs to the field of rock mechanics and geotechnical engineering. The method comprises the following steps: firstly, collecting surface structure characteristics of a rock material, and acquiring geometric characteristics of the surface of the rock material through a scanning technology; the microscopic structure of the rock material is modeled, a Voronoi division method is adopted to divide a numerical model, and in combination with surface feature data, geometric features of the rock material are precisely restored, model parameters are adjusted, and guarantee is provided for subsequent loading operation. And secondly, carrying out loading operation on the established model, simulating deformation and damage processes of the rock under different stress conditions, and monitoring a displacement field on the surface of the model in the loading process to obtain high-precision displacement data. And finally, judging the type of the destruction mode according to the displacement data. According to the method, the type of the rock numerical simulation microscopic crack can be judged, the method is convenient, and the result is accurate.

Description

Technical Field

[0001] The present invention belongs to the field of rock mechanics and geotechnical engineering, and relates to a method for solving transgranular or intergranular failure of a numerical model considering the mesoscopic structure of rocks. In particular, when the discrete element method cannot consider the strain components caused by block separation, the two-dimensional digital image correlation technology can solve the approximate displacement and strain conditions in a specified area, and determine the mesoscopic failure type of the model through the displacement and strain conditions thereof. Background Art

[0002] The rapid development of computers has brought efficient and low-cost tools to rock mechanics, and the numerical simulation method has become an indispensable part of the study of rock mechanics. The discrete element method regards the particles or blocks in rock-like materials as independent individuals, calculates the interaction forces between each individual, and obtains the whole process of mesoscopic fracture of rock-like materials. Since most real rock-like materials are composed of multiple minerals and the particle sizes of each mineral are uneven, the discrete element method is more suitable for describing this situation compared with other methods.

[0003] At present, some discrete element software only has the built-in function of extracting strain for the block area, and cannot consider the strain components caused by block separation. At present, there is still no mature method to solve this problem.

[0004] The two-dimensional digital image correlation technology is a visual measurement technology widely used for the deformation detection of rigid body structures. It is often used in the tests of real rock-like materials. By projecting or drawing a random speckle pattern on the surface of the object to be measured, defining the similarity function of the image, and analyzing the two images before and after the deformation of the object, the displacement field of the sampling points can be obtained, and thus the deformation can be obtained. In the tests of real rock-like materials, four requirements need to be met for using the two-dimensional digital image correlation technology for damage location: (1) A random speckle pattern is required on the surface of the sample; (2) The speckle size should be small, and the gray intensity values should have a relatively uniform distribution histogram; (3) The axis of the camera used to capture the image must be parallel to the surface normal of the sample; (4) The camera should be placed far enough so that the displacement measurement error caused by out-of-plane deformation is minimized.

[0005] Therefore, there is an urgent need for a new method that can obtain the displacement field and strain field generated during the failure process of a numerical model based on mesoscopic structure modeling, and further determine the mesoscopic failure type of rock-like materials, laying a foundation for further research on the mesoscopic failure of rock-like materials. Summary of the Invention

[0006] In view of the above problems existing in the prior art, the present invention provides a method for identifying mesoscopic failure types of numerical models of rock-like materials based on two-dimensional digital image correlation. The present invention applies the two-dimensional digital image correlation method to the displacement field generated during the failure process of a numerical model built based on a mesoscopic structure. The result of this method is intuitive and clear, which can accurately express the displacement field generated by the numerical model built based on the mesoscopic structure of rock-like materials, and further identify the types of mesoscopic cracks.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A method for solving transgranular or intergranular failure of a numerical model of rock-like materials based on two-dimensional digital image correlation. First, a numerical model is established based on the surface structure characteristics of rock-like materials, and the geometric characteristics of the surface of rock-like materials (including geometric morphology characteristics and mesoscopic structure information) are obtained through scanning technology; the mesoscopic structure of rock-like materials is modeled, and the numerical model is divided by the Voronoi partitioning method. Combining the surface feature data, the geometric characteristics of rock-like materials are accurately restored, and the model parameters are adjusted to provide guarantee for subsequent loading operations. Secondly, the established model is subjected to loading operations to simulate the deformation and failure processes of rocks under different stress conditions. During the loading process, the two-dimensional digital image correlation (2D-DIC) method is used to monitor the displacement field on the surface of the model to obtain high-precision displacement data. Finally, the type of failure mode (such as transgranular failure or intergranular failure) is identified according to the displacement data. Specifically, it includes the following steps:

[0009] Step 1: Collect the surface structure feature data of real rock-like materials, and obtain the geometric morphology characteristics and mesoscopic structure information of the surface of rock-like materials through high-precision scanning technologies such as Canon cameras. The collected image data is subjected to gray-scale processing to extract the gray-scale distribution information on the surface of rock-like materials. Based on the gray-scale processed image data, the discrete element method (DEM) is used to model the mesoscopic structure of rock-like materials. Combining the gray-scale data, the geometric characteristics of rock-like materials are accurately restored, and the numerical model is divided by the Voronoi partitioning method to ensure that the model can truly reflect the mesoscopic structure characteristics of rocks. By adjusting the model parameters (such as contact characteristics and material properties), the accuracy of the numerical model is optimized, providing a reliable basis for subsequent mechanical behavior analysis and failure mode research. Specifically as follows:

[0010] Step 1.1, Use high-precision scanning technology to collect the surface image data of rock-like materials, and perform gray-scale processing on the collected image data. The formula for gray-scale processing is:

[0011] Gray(i,j) = 0.299 * R(i,j) + 0.578 * G(i,j) + 0.114 * B(i,j)

[0012] Among them, Gray(i, j) represents the gray value of the pixel at the i-th row and j-th column in the image, R(i, j) represents the red channel intensity value of the pixel at the i-th row and j-th column in the image, G(i, j) represents the green channel intensity value of the pixel at the i-th row and j-th column in the image, and B(i, j) represents the blue channel intensity value of the pixel at the i-th row and j-th column in the image. A gray image of the surface of the rock-like material is obtained.

[0013] Step 1.2: Based on the gray image of the surface of the rock-like material obtained in Step 1.1, a numerical model is established using the discrete element method. The numerical model is divided using the Voronoi partitioning method. A set of generation points P = {p 1 , p 2 , …, p n} is given in the numerical model. The numerical model is divided into n regions V = {V 1 , V 2 , …, V n} using Voronoi partitioning. Each region V i is defined as:

[0014]

[0015] where x is an arbitrary point in the numerical model, R d is a region in the numerical model; d(x, p i ) is the distance between point x and the generation point p i ; p i is the i-th generation point in the numerical model; d(x, p j ) is the distance between point x and the generation point p j ; p j is the j-th generation point in the numerical model; V i is called the Voronoi cell of the generation point p i .

[0016] Step 1.3: According to the mineral composition of the rock-like material and the indoor experimental data of the rock-like material, material parameters such as the elastic modulus, Poisson's ratio, and tensile strength of the material, as well as the contact parameters between minerals in the rock-like material, are defined. By comparing the numerical simulation results with the indoor experimental data, the accuracy of the model is verified. If the difference is large, the parameters are readjusted until the simulation results are in good agreement with the experimental data.

[0017] Step 2: Perform a loading operation on the numerical model obtained in Step 1 to simulate the deformation and failure process of the rock under different stress conditions. During the loading process, the two-dimensional digital image correlation (2D-DIC) method is used to monitor the displacement field on the surface of the numerical model to obtain high-precision displacement data. Specifically as follows:

[0018] Step 2.1, Image acquisition: Before loading the numerical model obtained in Step 1, the surface of the numerical model is acquired as a reference image (undeformed state). During the loading process, m deformed images (deformed state) of the surface of the numerical model are acquired at certain time intervals or load steps. A total of one reference image (undeformed state) and m deformed images (deformed state) are obtained.

[0019] Step 2.2, Select the region of interest (ROI) for the reference image (undeformed state) and the deformed images (deformed state), and divide each region of interest (ROI) of the reference image and the deformed images into a sub-regions in the same method. Each sub-region contains a certain number of pixel points. In this step, the number of finally divided sub-regions of the reference image and the deformed images is the same (the number of sub-regions is the same, but the positions are uncertain. The corresponding sub-regions of the reference image and the deformed images need to be found according to the similarity criterion).

[0020] Step 2.3, Use the correlation function to measure the similarity between the sub-regions of the reference image and the sub-regions of the deformed image;

[0021] The correlation function refers to the zero-mean normalized cross-correlation function (ZNCC), as follows:

[0022]

[0023] where f(x, y) is the gray value of the sub-region of the reference image; g(x′, y′) is the gray value of the sub-region of the deformed image; f avg and g avg are the average gray values of the sub-region of the reference image and the sub-region of the deformed image respectively. By calculating the correlation of each sub-region in the reference image and the deformed image, the corresponding relationship between the two is determined; when the correlation value is 1, the similarity is the highest, indicating that these two sub-regions are completely corresponding.

[0024] In each sub-region of the deformed image, compared with the sub-region of the reference image, use the correlation function to calculate the similarity of the pixel points in each sub-region of the deformed image, find the position where the correlation function value is the largest, and record the displacement vector (u, v) of this sub-region, where u is the horizontal displacement and v is the vertical displacement. Each sub-region corresponds to a displacement vector.

[0025] Step 2.4, Combine the displacement vectors of all sub-regions of the deformed image to form a displacement field matrix. The displacement field matrix contains the horizontal displacement u and the vertical displacement v of each pixel point in the deformed image. The displacement field matrices of m deformed images are obtained.

[0026] Step 3: Determine the type of failure mode (such as transgranular failure or intergranular failure) according to the displacement field matrix. The specific steps are as follows:

[0027] Step 3.1, group the reference images according to the types of rock material components to obtain the image data after grouping the reference images. Specifically:

[0028] If the rock material is composed of b mineral materials, determine the colors of the mineral materials, divide the gray values into b material intervals according to the colors of each mineral material, and each material interval represents a mineral material.

[0029] Step 3.2, process the displacement field matrix of the m deformed images obtained in Step 2. In the displacement field matrix, find the maximum points for the horizontal displacement and vertical displacement data respectively, and consider the maximum points as the crack generation points. Obtain the images of m crack generation points. That is, one deformed image yields one displacement field matrix, and one displacement field matrix yields one crack generation point image.

[0030] Step 3.3, compare the image data of the m crack generation points obtained in Step 3.2 with the image data of the reference images after grouping obtained in Step 3.1: If the crack generation point appears in the nth material interval, it is considered that the crack is a transgranular crack of this material; if the crack generation point appears at the junction of multiple material intervals, it is considered that the crack is an intergranular crack.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0032] (1) The present invention applies the two-dimensional digital image correlation technology commonly used in indoor experiments to the discrimination of crack types in numerical model images.

[0033] (2) The present invention can accurately express the displacement field generated by the numerical model established based on the mesoscopic structure of rock materials.

[0034] (3) The present invention has significant application value and can effectively discriminate the mesoscopic crack types generated during the failure process of rock materials, laying a foundation for revealing the evolution law of failure cracks in rock materials. Description of the Drawings

[0035] Figure 1 It is a flow chart for discriminating the failure types of cracks in the numerical model of rock materials by the two-dimensional digital image correlation provided by the present invention.

[0036] Figure 2 It is the gray-scale processing of the surface image data of fine-grained granite by the present invention.

[0037] Figure 3 It is the surface image data of the numerical model: (a) the surface image (reference image) before loading the numerical model, (b) the image at the time of failure of the numerical model (deformed image).

[0038] Figure 4Displacement nephogram of the numerical model: (a) Horizontal displacement nephogram; (b) Vertical displacement nephogram.

[0039] Figure 5 Schematic diagrams of transgranular cracks and intergranular cracks.

[0040] Figure 6 Curve of the number of mesoscopic cracks and strain of fine-grained granite. Specific implementation manners

[0041] In order to further explain the technical solution of the present invention, the present invention will be elaborated in detail below with reference to the accompanying drawings and embodiments.

[0042] As Figure 1 shown, the specific process of the two-dimensional digital image correlation method includes the following steps:

[0043] Step 1: In this embodiment, fine-grained granite with a height of 100 mm and a width of 50 mm is taken as an example for detailed elaboration. Collect the surface structure feature data of the fine-grained granite, and use a Canon camera to obtain the geometric morphology features and mesoscopic structure information of the surface of the rock-like material. Perform gray-scale processing on the collected image data, and extract the gray-scale distribution information on the surface of the rock-like material. Based on the gray-scale processed image data, use UDEC to model the mesoscopic structure of the rock-like material, accurately restore the geometric features of the rock-like material in combination with the gray-scale data, and use the Voronoi division method to divide the numerical model to ensure that the model can truly reflect the mesoscopic structure characteristics of the rock. By adjusting the model parameters (such as contact characteristics and material properties), optimize the accuracy of the numerical model, and provide a reliable basis for subsequent mechanical behavior analysis and failure mode research. Specifically as follows:

[0044] Step 1.1, use a Canon camera to collect the surface image data of the rock-like material, and perform gray-scale processing on the collected image data (such as Figure 2 ), and the formula for gray-scale processing is:

[0045] Gray(i,j) = 0.299*R(i,j) + 0.578*G(i,j) + 0.114*B(i,j)

[0046] Among them, Gray(i, j) represents the gray value of the pixel at the i-th row and j-th column in the image, R(i, j) represents the red channel intensity value of the pixel at the i-th row and j-th column in the image, G(i, j) represents the green channel intensity value of the pixel at the i-th row and j-th column in the image, and B(i, j) represents the blue channel intensity value of the pixel at the i-th row and j-th column in the image. In this embodiment, the pixel range of the image is: i ∈ (0, 2300), j ∈ (0, 1150). According to the proportion of mineral components of fine-grained granite, the selected thresholds in this embodiment are 63 and 127 respectively. The gray value between 0 and 63 is considered to be mica mineral, the gray value between 63 and 127 is considered to be albite mineral, and the gray value between 127 and 255 is considered to be quartz mineral.

[0047] Step 1.2: Based on the gray-scale image of the surface of the rock-like material obtained in Step 1.1, establish a numerical model using the discrete element method. Divide the numerical model using the Voronoi partitioning method, and given a set of generating points P = {p 1 , p 2 , …, p n} in the numerical model. The Voronoi partitioning divides the space into n regions V = {V 1 , V 2 , …, V n}, where each region V i is defined as:

[0048]

[0049] where x is an arbitrary point in the numerical model, R d is a region in the numerical model; d(x, p i ) is the distance between point x and the generating point p i ; p i is the i-th generating point in the numerical model; d(x, p j ) is the distance between point x and the generating point p j ; p j is the j-th generating point in the numerical model; V i is called the Voronoi cell of the generating point p i . In this embodiment, the average side length of the Voronoi cell is defined as 2 mm.

[0050] Step 1.3: Define material parameters such as the elastic modulus, Poisson's ratio, and tensile strength of the material, as well as the contact parameters between minerals of the rock-like material, according to the mineral composition of the rock-like material and the indoor experimental data of the rock-like material. Verify the accuracy of the model by comparing the numerical simulation results with the indoor experimental data. If the difference is large, readjust the parameters until the simulation results are in good agreement with the experimental data. The optimized parameters obtained in this embodiment are as follows:

[0051] Table 1 Physical and Mechanical Parameters of Mineral Grains in Fine-Grained Granite

[0052]

[0053] Table 2 Microscopic Contact Parameters of Fine-Grained Granite

[0054]

[0055]

[0056] Note: k n is the contact normal stiffness, k s is the contact shear stiffness, C is the cohesion, is the internal friction angle, and is the tensile strength

[0057] Step 2: Perform a loading operation on the numerical model obtained in Step 1 to simulate the deformation and failure processes of the rock under different stress conditions. During the loading process, use the two-dimensional digital image correlation (2D-DIC) method to monitor the displacement field on the model surface and obtain high-precision displacement data. Specifically as follows:

[0058] Step 2.1, Image acquisition: Before loading the numerical model obtained in Step 1, acquire the surface of the numerical model as a reference image (undeformed state) ( Figure 3 (a)), and during the loading process, at a certain time interval or load step, acquire the deformed images (deformed state) of the surface of the numerical model. In this embodiment, 1 deformed image is obtained, and the image at the time of failure of the numerical model is selected as the deformed image ( Figure 3 (b)).

[0059] Step 2.2, In this embodiment, select the region of interest (ROI) for the reference image (undeformed state) and the deformed image (deformed state), set the entire numerical model as the region of interest, and divide each region of interest (ROI) of the reference image and the deformed image into 2,596,896 sub-regions in the same method, and set the radius of the sub-regions to 15.

[0060] Step 2.3, Use the correlation function to measure the similarity between the sub-regions of the reference image and the sub-regions of the deformed image;

[0061] The correlation function refers to the zero-mean normalized cross-correlation function (ZNCC), as follows:

[0062]

[0063] where f(x, y) is the gray value of the sub-region of the reference image, g(x′, y′) is the gray value of the sub-region of the deformed image, f avg and g avgThey are the average gray values of the reference image sub-region and the deformed image sub-region respectively. By calculating the correlation between each sub-region in the reference image and the deformed image, the corresponding relationship between the two is determined; when the correlation value is 1, the similarity is the highest, indicating that these two sub-regions correspond exactly.

[0064] In each sub-region of the deformed image, comparing with the sub-region of the reference image, the similarity of each sub-region in the deformed image is calculated using the correlation function, the position where the correlation function value is the largest is found, and the displacement vector (u, v) of this sub-region is recorded, where u is the horizontal displacement and v is the vertical displacement. Each sub-region corresponds to a displacement vector.

[0065] Step 2.4, Combine the displacement vectors of all sub-regions of the deformed image to form a displacement field matrix. The displacement field matrix contains the horizontal displacement u and the vertical displacement v of each pixel point in the deformed image. In this embodiment, 1 displacement field matrix of the deformed image is obtained, and the displacement field matrix is represented by color mapping as Figure 4 .

[0066] Step 3: Determine the type of failure mode (such as transgranular failure or intergranular failure) according to the displacement field matrix. The specific steps are as follows:

[0067] Step 3.1, Group the reference image according to the types of fine-grained granite components to obtain the image data after grouping the reference image. Specifically:

[0068] For example, if the fine-grained granite is composed of 3 mineral materials (mica, albite, and quartz respectively), determine the colors of the mineral materials. Among them, the color of quartz is relatively white, and the color of mica is relatively black. Therefore, it is considered that the mica mineral has a gray value between 0 - 63, the albite mineral has a gray value between 63 - 127, and the quartz mineral has a gray value between 127 - 255. Divide it into 3 material intervals according to each mineral. Each material interval represents a mineral material.

[0069] Step 3.2, Process the displacement field matrix of the 1 deformed image obtained in Step 2. In the displacement field matrix, find the maximum value points for the horizontal displacement and vertical displacement data respectively, and consider the maximum value points as the crack generation points. An image of the crack generation points is obtained.

[0070] Step 3.3, Compare the image data of the 1 crack generation points obtained in Step 3.2 with the image data of the reference image after grouping obtained in Step 3.1: If the crack generation points appear in the first material interval, it is considered that the crack is a transgranular crack of the mica mineral material; if the crack generation points appear at the junction of multiple material intervals, it is considered that the crack is an intergranular crack. In this embodiment, 7 situations may occur, where situations 1 - 3 are transgranular failures, and situations 4 - 7 are intergranular failures (such as Figure 5(For example: Case 1 is a transgranular crack in the mica mineral, Case 4 is an intergranular crack at the contact between the mica mineral and the albite mineral, and the same applies to other cases) as shown in the following table.

[0071] Table 3 Possible cases of crack data

[0072]

[0073] Note: 1 indicates that there is a crack at this pixel point, and 0 indicates that there is no crack at this pixel point.

[0074] In this embodiment, the mesoscopic cracks in the whole process of loading and failure of fine-grained granite are statistically analyzed according to the above method, and the relationship between the number of transgranular cracks and intergranular cracks in fine-grained granite and strain is as Figure 6 .

[0075] The above-described embodiments only represent the implementation manners of the present invention, but should not be construed as limiting the scope of the present invention. It should be noted that for those skilled in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation, characterized in that: The rock material numerical model mesoscopic failure type identification method comprises the following steps: Step 1: Collect surface structural feature data of real rock materials, and obtain the geometric features of the surface of rock materials through scanning technology, including geometric morphology and microstructure information; grayscale process the collected image data to extract the grayscale distribution information of the surface of rock materials; model the microstructure of rock materials based on the grayscale processed image data, accurately restore the geometric features of rock materials in combination with grayscale data, divide the numerical model, and ensure that the numerical model can truly reflect the microstructure characteristics of rocks; optimize the accuracy of the numerical model by adjusting the model parameters; Step 2: Perform loading calculations on the numerical model obtained in step 1 to simulate the deformation and failure process of rock under different stress conditions; during the loading process, monitor the displacement field on the surface of the numerical model to obtain high-precision displacement data; Step 3: Determine the type of failure mode based on the displacement field matrix.

2. According to claim 1, a method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation is characterized in that: The step 1 is specifically as follows: Step 1.1, collecting image data of the surface of rock materials, performing grayscale processing on the collected image data, and obtaining a grayscale image of the surface of the rock materials; Step 1.2, based on the grayscale image of the rock material surface obtained in step 1.1, a numerical model is established using the discrete element method; the numerical model is divided using the Voronoi partitioning method, and a set of generating points P = {p1, p2, ..., p n }, the numerical model is divided into n regions V = {V1, V2, ..., V n }, where each region V i Defined as: Where x is any point in the numerical model, R d is a region in the numerical model; d(x, p i ) is the point x and the generating point p i The distance between i is the i-th generating point in the numerical model; d(x, p j ) is the point x and the generating point p j The distance between j is the jth generation point in the numerical model; V i Generator point p i Voronoi cells; Step 1.3, simulate the numerical model and adjust the parameters to ensure the accuracy of the numerical model.

3. The method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation according to claim 2 is characterized in that: In step 1.1, the formula for grayscale processing is: Gray(i,j)=0.299*R(i,j)+0.578*G(i,j)+0.114*B(i,j) Among them, Gray(i,j) represents the grayscale value of the pixel in the i-th row and j-th column of the image, R(i,j) represents the red channel intensity value of the pixel in the i-th row and j-th column of the image, G(i,j) represents the green channel intensity value of the pixel in the i-th row and j-th column of the image, and B(i,j) represents the blue channel intensity value of the pixel in the i-th row and j-th column of the image.

4. The method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation according to claim 2 is characterized in that: The step 1.3 is specifically as follows: based on the mineral composition of the rock material and the experimental data of the rock material, define the elastic modulus, Poisson's ratio and tensile strength material parameters of the material, as well as the contact parameters between the minerals of the rock material; verify the accuracy of the model by comparing the numerical simulation results with the experimental data; if the difference is large, readjust the parameters until the simulation results are highly consistent with the experimental data.

5. The method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation according to claim 1 is characterized in that: The step 2 is specifically as follows: Step 2.1, image acquisition: before loading the numerical model obtained in step 1, collect the surface of the numerical model as a reference image to represent the undeformed state; during the loading process, collect m deformation images of the surface of the numerical model at a certain time interval or load step to represent the deformation state; A total of one reference image and m deformed images are obtained; Step 2.2, selecting a region of interest ROI for the reference image and the deformed image, and dividing each region of interest ROI of the reference image and the deformed image into a sub-regions according to the same method, each sub-region contains a certain number of pixels; in this step, the reference image and the deformed image are finally divided into the same number of sub-regions; Step 2.3, use the correlation function to measure the similarity between the reference image sub-region and the deformed image sub-region; specifically: In each sub-region of the deformed image, the similarity of the pixels in each sub-region of the deformed image is calculated by using the correlation function in comparison with the sub-region of the reference image, the position where the correlation function value is the largest is found, and the displacement vector (u, v) of the sub-region is recorded, where u is the horizontal displacement and v is the vertical displacement; each sub-region corresponds to a displacement vector; Step 2.4, combining the displacement vectors of all sub-areas of the deformed image to form a displacement field matrix; the displacement field matrix includes the horizontal displacement u and vertical displacement v of each pixel in the deformed image; and obtaining the displacement field matrices of m deformed images.

6. The method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation according to claim 5 is characterized in that: In step 2.3, the correlation function refers to the zero-mean normalized cross-correlation function, as follows: Where f(x, y) is the grayscale value of the sub-region of the reference image; g(x′, y′) is the grayscale value of the sub-region of the deformed image; f avg and g avg are the average grayscale values ​​of the reference image sub-region and the deformed image sub-region respectively; by calculating the correlation between the reference image and each sub-region in the deformed image, the correspondence between the two is determined; when the correlation value is 1, the similarity is the highest, indicating that the two sub-regions completely correspond.

7. According to the method for identifying mesoscopic failure types of rock material numerical models based on two-dimensional digital image correlation according to claim 5, the step 3 is specifically: Step 3.1, grouping the reference images according to the types of rock material components to obtain image data after the reference images are grouped; Step 3.2, processing the displacement field matrix of the m deformed images obtained in step 2, finding the maximum point of the horizontal displacement and vertical displacement data in the displacement field matrix, and considering the maximum point as the crack initiation point; obtaining images of the m crack initiation points; Step 3.3, compare the image data of the m crack initiation points obtained in step 3.2 with the image data after the reference image grouping obtained in step 3.1: if the crack initiation point appears in the nth material interval, it is considered that the crack is a transgranular crack of the material; if the crack initiation point appears at the junction of multiple material intervals, it is considered that the crack is an intergranular crack.

8. According to the method for identifying the micro-destruction type of a numerical model of rock materials based on two-dimensional digital image correlation according to claim 7, step 3.1 is specifically: if the rock material is composed of b types of mineral materials, determine the color of the mineral material, and divide the grayscale value into b material intervals according to the color of each mineral material, and each material interval represents a mineral material.

Images